Computer-based problem-solving is fast becoming the accepted paradigm for testing students and practitioners on national and specialty board examinations. Computer-based learning of diagnostic skill has not yet reached the same level of acceptance. A microcomputer-based expert system (iliad) has been developed to model the way experts in each medical subspecialty form hypotheses and decide which data to acquire at each stage of a patient workup. The impact of this system on the problem-solving ability of medical students during their clerkship on the medical service o the three teaching hospitals of the University of Utah will be evaluated by comparing each student's performance as cases they have and have not worked up as simulations using Iliad.